Back at Veginova to build the data, BI, and reporting the business runs on. Extends the reporting stack I owned in 2025 into the finance and production layer.

• Built the finance source of truth in Power BI and PostgreSQL by modelling invoice lines into a semantic model reconciled line by line to the official accounts, making profit per product and per customer visible for the first time.
• Built the production planning engine in SQL that sets what to produce, how much, and when on a one-year seed lead time. It reproduces the planner's own numbers exactly and tests a scenario before seed is committed.
• Shipped forecasting into the planning model with Python and SQL views, so next season's production need updates on every sale instead of going stale in a spreadsheet.
• Designed the Microsoft Fabric migration path for the reporting stack (Dataflows Gen2 ingestion, warehouse models, notebook forecasting, Direct Lake semantic models), with the current reconciliation totals as the acceptance test.

Stack: Power BI, SQL, PostgreSQL, Python, Supabase, Claude Code

Owned the data work behind a Danish SMB accounting and fintech firm: ICP, scoring, software detection, and full-channel revenue analysis, from raw source to scores feeding sales.

• Built an ML classifier in scikit-learn on Playwright-scraped features that identified a prospect's accounting software across the Danish SMB market at holdout AUC 0.75 (permutation test p < 0.0001), surfacing high-fit prospects on competing platforms.
• Broke down close rate, conversion, meeting time, and pipeline velocity across five channels (Meta, Google, LinkedIn outbound, cold calling, inbound) in SQL and Python, identifying which channels drove customers and which burned budget.
• Built ICP and predictive lead-scoring models on CRM and Danish business-registry data that cut low-quality leads reaching sales and sharpened outbound prioritisation.
• Rebuilt the sales commission model with forecasting weightings toward subscription and MRR, removing the incentive to discount monthly deals to close.
• Shipped the scores into an end-to-end LinkedIn pipeline (scraping, enrichment, scoring, sequencing) with ~50% cold acceptance.

Stack: SQL, Python, Power BI, Pipedrive, Clay, Playwright, Apify, HeyReach, Claude Code

Promoted into BI ownership: the data infrastructure, Power BI dashboards, and analytics across sales, operations, and production.

• Improved KPI accuracy by 40%+ by standardising metric definitions in the semantic model across sales, operations, and production and fixing the source data underneath the dashboards.
• Cut reporting time by 10+ hours per week by replacing weekly manual rebuilds with repeatable SQL and Power BI pipelines.
• Owned the single source of truth: Power BI dashboards, SQL data models, and automated reporting pipelines across the business.
• Resolved data inconsistencies across ERP, CRM, and operational tools by finding where the flows broke and rebuilding them with reliable source logic.

Stack: SQL, Power BI, DAX, Python, Excel, Supabase, Claude Code

Started in marketing and lead sourcing, moved into RevOps. Built the operational data layer between sales, marketing, and production that grew into the Data Analyst role.

• Built automated data collection and enrichment pipelines in Clay and n8n for prospecting, replacing manual spreadsheet work.
• Connected ERP, CRM, and operational tools into reliable cross-functional data flows, and built the reporting leadership used for weekly decisions.

Stack: SQL, Power BI, Clay, n8n, Supabase, Apify, Excel

Service business building AI-driven lead generation, enrichment, and CRM automation for B2B companies. It closed without customers; it's where I learned which automation is worth building.

• Built lead enrichment and scoring workflows wiring collection, enrichment, and qualification into one pipeline, improving qualification accuracy by 30–40%.
• Integrated HubSpot and GoHighLevel with external data sources and outreach tools so prospects flowed from collection to sales-ready.

Stack: SQL, Python, Clay, n8n, Apify, HubSpot, GoHighLevel, HeyReach

• Reviewed and annotated 1,000+ Danish-language AI conversations, feeding model failure patterns back into training data. The same understanding of where LLMs break powers the AI scoring workflows I build today.

• Managed product and marketing data in ERP and CRM platforms across European markets and built Excel dashboards for campaign reporting.